Post #288998

Yet another reminder that "chatbot" AIs are programmed to give something that looks like the most likely answer based on text in their training data. They have no understanding of the real world or what their words mean. (PS, note the follow-up clarification: no one asked the Slack bot about the fire alarm, since no one expected it to have relevant knowledge. But still, it eagerly butted into a conversation between evacuated staff.) https://mastodon.online/@tagir_valeev/116057271527521893